Construction and Building Materials 203 (2019) 501–513

Contents lists available at ScienceDirect

Construction and Building Materials

journal homepage: www.elsevier.com/locate/conbuildmat

Strength development and microstructural characteristics of soft

dredged clay stabilized with basic oxygen furnace steel slag
Gyeongo Kang a, Arlyn Aristo Cikmit b, Takashi Tsuchida b,⇑, Hideki Honda c, Young-sang Kim d
a
Honam Regional Infrastructure Technology Management Center, Chonnam National University, 50 Daehak-ro, Yeosu, Jeollanamdo 550-749, Republic of Korea
b
Department of Civil and Environmental Engineering, Graduate School of Engineering, Hiroshima University, 1-4-1, Kagamiyama, Higashi-Hiroshima, Hiroshima 739-8527, Japan
c
Steel Research Laboratory, JFE Steel Corporation, 1-1, Minami Watarida-Cho, Kawasaki City, Kanagawa 210-0855, Japan
d
Department of Civil Engineering, Chonnam National University, 77 Yongbong-ro, Buk-gu, Gwangju 61186, Republic of Korea

h i g h l i g h t s

Strength development of BOS-treated clay were studied from immediately after mixing.

Strength development of BOS-treated clay can be classified into three zones.

Cementitious compounds is found in BOS-treated clay.

BOS is feasible as a binder to replace ordinary Portland cement.

a r t i c l e i n f o a b s t r a c t

Article history: This paper presents the results of the strength development and macrostructure properties of soft
Received 31 January 2018 dredged clay stabilized with basic oxygen furnace steel (BOS) slag, using the vane shear test, the uncon-
Received in revised form 17 August 2018 fined compression test, and scanning electron microscope. The results indicate that the flowability exhib-
Accepted 17 January 2019
ited a strong correlation with the undrained shear strength. A dormant period was observed, during
which the BOS-treated clay strength did not increase during the 0.5–2 h curing times. Moreover, the
strength development can be classified into three zones according to the strength increment rate in
Keywords:
the log-log scale graph: 1) inactive zone (0.5–5 h), 2) active zone (5 h to 3 d), and 3) moderate zone
Basic oxygen furnace steel slag
Soft dredged clay
(3–90 d). Empirical equations for each zone were proposed based on the volumetric solid content, initial
Unconfined compressive strength water content, strength at 1 h of curing, and strength increment coefficients. Finally, the reticulation
Ground improvement structure of the amorphous C-S-H gel and platy AFm phases on all samples were observed, with inter-
Reclamation growths of rod-like ettringite within the flat clay structure and flocculated clay-cement cluster, and
the cementitious compounds became more evident with an increase in the BOS content and curing time.
Ó 2019 Published by Elsevier Ltd.

1. Introduction clays stabilized with binders such as ordinary Portland cement

A large amount of soft clayey soils at the sea bottom is dredged
annually to maintain ports and secure navigable waterways. As
certain difficulties and economic issues exist in securing suitable
disposal sites for dredged deposits, alternative green and econom-
ical approaches for recycling and applying dredged clay have
become a major research topic. In general, dredged soil exhibits
extremely low shear strength, high fine content, high compressibil-
ity, and a natural water content that is commonly higher than its
liquid limit [17]. As a result, dredged clayey soils have never been
reused as raw materials without any special treatment. Dredged
⇑ Corresponding author.
E-mail address: ttuchida@hiroshima-u.ac.jp (T. Tsuchida).
and lime have been studied extensively studied and implemented
as construction materials, filling and reclamation materials, artifi-
cial barrier layers, and submerged embankments in ports and air-
port areas [34,39,41].
Steel slag is a by-product material generated during steelmak-
ing and refining operations in the steel manufacturing process.
Steel slags can be classified into two types, namely basic oxygen
furnace steel (BOS) slag and electric arc furnace (EFA) slag.
Previous studies have demonstrated that steel slag exhibits supe-
rior engineering characteristics, such as a high friction angle, low
water absorption, high stiffness, abrasion resistance, and high com-
pactibility. These properties are the main reason for its extensive
utilization in road construction, ground improvement (sand com-
paction pile methods), and asphalt mixtures as aggregate materials

https://doi.org/10.1016/j.conbuildmat.2019.01.106
0950-0618/Ó 2019 Published by Elsevier Ltd.
502 G. Kang et al. / Construction and Building Materials 203 (2019) 501–513
Notations
BOS basic oxygen furnace steel slag
BOSvol (%) ratio of BOS volume to total volumes
FA flowability
Y volumetric solid content
w0/wLL initial water content
BOSmass ratio of BOS mass to mass of soil and BOS particles
t curing time period
b1 strength increment coefficient in inactive zone (gradient
for relationship between strength and curing time for
0.5–5 h of curing)
[4,43]. However, it is well known that the steel slag contains inac-
tive materials referred to as free-CaO and free-MaO, which can
hydrate and nearly double their volume through hydration, and
their expansion potential with time causes slag cracking [30,9].
An aging treatment process is subsequently required to prevent
volume expansion prior to utilization as aggregate materials.
BOS consists of dicalcium silicate (C2S), the rhombohedral to
orthorhombic (R-O) phase (solid solution of CaO, FeO, MgO, and
MnO), tricalcium silicate (C3S), tetracalcium aluminoferrite
(C4AF), dicalcium ferrite (C2F), free-lime (f-CaO), olivine, and mer-
winite, and its chemical and mineralogy properties are similar to
those of Portland cement [21,5,26,6]. BOS has been referred to as
a weak Portland cement, owing to its low C3S and C2S contents
[18,32,25]. According to the BOS mineral components, it can be
used as a stabilizer with or without an activator to solidify soils
instead of other binders, as well as an aggregate in civil engineer-
ing works. Recently, in Japan, dredged clayey soils stabilized with
BOS have been used as various geo-materials requiring relatively
low strength, such as the filling material in quay walls, submerged
embankments, and breakwater [7,37]. Moreover, the environmen-
tal impact was confirmed, in order to successfully reuse BOS as a
soil stabilizer and filling material. The safety assessment for envi-
ronmental impact was divided into two categories; 1) safety
against harmful substances, such as heavy metals and toxic sub-
stances, and 2) safety against pH value. The safety evaluation for
harmful substances was conducted following the Act on the
Prevention of Marine Pollution and Marine Accidents in Japan
[1]. The pH effect of environmental safety was determined by
means of a test method for the pH of suspended soils [10], whereby
the pH value was monitored on the site and compared to the
results of laboratory experiments. According to the results, the
BOS-treated dredged clay met all of the criteria for stability against
environmental impact, and also reaffirmed the environmental
impact stability by means of monitoring and testing in the actual
field [7]. This application method may provide an environmentally
friendly manner of recycling dredged clayey soils as construction
materials, and expanding the use of steel slag as a binder on behalf
of cement that emits anthropogenic CO2. Over the past several dec-
ades, the majority of the research regarding steel slag has focused
on road base course material [31], ground improvement material
[35,16], aggregates in asphalt mixtures [2,24], concrete aggregates
[3,27], and cement additives [19,26]. Based on the literature
review, the study of BOS commonly focuses on its applicability
as an aggregate material in diverse fields. Consequently, studies
investigating the utilization of BOS for soft dredged clay stabiliza-
tion are very scarce in countries other than Japan [25,6]. Further-
more, the mechanism of the strength mobilization of dredged
clay treated with BOS over the curing time immediately following
mixing has also not yet been clarified.
The main objective of this study is to investigate the strength
development of dredged clay stabilized by BOS at various curing
a1 strength parameter (strength at 1 h of curing)
b2 strength increment coefficient in active zone (gradient
for relationship between strength and curing time for
5–3 d of curing)
v0 /vLL specific volume normalized with that at liquid limit
c1 and c2 parameters depending on specific volume ratio
b3 strength increment coefficient in active zone (gradient
for relationship between strength and curing time for
3–90 d of curing)
times in order for it to be used as a geo-material in geotechnical
applications, while considering its microstructural and mineralog-
ical properties. In order to evaluate the stabilized dredged clay
strength development, a series of laboratory vane shear (LVS)
and unconfined compression (UC) tests were carried out on speci-
mens prepared with different BOS contents, as well as the water
content at various curing times immediately following mixing,
ranging from 0.5 h to 90 d. Moreover, the microstructural proper-
ties of the dredged clay stabilized by BOS were analyzed by means
of a scanning electron microscope (SEM) and an energy dispersive
spectroscopy (EDS) analysis.
2. Experimental program
2.1. Materials
Marine clay dredged at the Tokuyama Port, Yamaguchi prefec-
ture, Japan, was used as the base clay in this experiment. The
geotechnical properties of the dredged marine clay are summa-
rized in Table 1. The liquid and plastic limits were 107.15% and
38.64%, respectively. This soil exhibits a high plasticity, with a plas-
tic index of 68.50%. The ignition loss was 8.17%, while the specific
gravity was 2.65. The marine dredged clay was composed of 9.98%
coarse-grained soils (>75 lm) and 90.02% fine-grained soils
(<75 lm). Based on the Unified Soil Classification System (USCS),
the dredged clay used in this study was classified as an inorganic
clay of high plasticity, or organic clay of medium to high plasticity
(CH-OH).
The BOS used in this experiment was a product of the
Fukuyama steel plant of the JFE Steel Corporation, and its physical
properties are listed in Table 2. The BOS particle sizes were less
than 5 mm, and the BOS contained a 6.48% moisture content of
its air-dried mass. The saturated surface and absolute dry densities
of the BOS were 3.15 and 3.02, respectively, while the water
absorption rate was 4.18%. The BOS free-CaO content was deter-
mined to dissolve an ethylene glycol among CaO using phenol
extraction and atomic absorption analysis, and its value was
Table 1
Physical properties of dredged clay at Tokuyama Port.
Property Value
Liquid limit, wLL (%) 107.15
Plastic limit, wPL (%) 38.64
Plasticity index, wPI (%) 68.50
Ignition loss, Li (%) 8.17
Specific gravity, Gs (g/cm3) 2.65
Coarse-grained soil (%) 9.98
Fine-grained soil (%) 90.02
Unified Soil Classification System (USCS) CH-OH
pH 7.2
 G. Kang et al. / Construction and Building Materials 203 (2019) 501–513 503

Table 2 Table 3
Basic properties of BOS. Chemical composition of BOS.

Property Value Chemical BOS (%) of this study BOS (%)a BOS (%)b
3
Saturated surface-dry density (g/cm ) 3.15 SiO2 14 8–20 8.6–13.1
Absolute dry density (g/cm3) 3.02 Al2O3 2.7 1–6 1.7–2.1
Water absorption rate (%) 4.18 FeO 33 10–35 28.3–32
Initial water content (%) 6.48 CaO 40 30–55 40.1–45
Particle size (mm) Less than 5 mm MgO 2.3 5–15 4.5–7.5
Coarse-grained soil (%) 99.50 MnO 4.1 2–8 2.0–4.1
Fine-grained soil (%) 0.50 TiO2 0.49 0.4–2 0.5–0.9
Free Calcium, F-CaO (%) 4.27 S 0.029 0.05–0.15 0.4–1.2
P 3.1 0.2–2 1.4–2.4
Cr 0.19 0.1–0.5 N/Ac
a
4.27% [38]. The particle-size distribution curves for the Tokuyama Shi [33].
b
dredged clay and BOS are plotted together in Fig. 1. Belhadj et al. [5].
c
Not available.
The chemical composition percentages of the BOS obtained by
X-ray fluorescence analysis and comparisons with the chemical
composition ranges for different BOS types obtained from the liter- Table 4
ature are provided in Table 3. In the BOS used in this study, cal- Type and condition of tests conducted in laboratory.
cium, iron, and silicon oxides accounted for more than 87%, and
Type of test Normalized initial BOS content, Curing time
the chemical composition ranges were very similar to those of water content, w0/wLL BOSVol (%)
the literature [33,5]. Furthermore, Belhadj et al. [5] reported that
LVS and UC* 1.2, 1.5, 20, 30 0.5, 2, 5, 7, 10, 15 (h)
the BOS composition agrees with that of clinker, except for the high 1.7, 2.0 1, 2, 3, 7, 28, 90 (days)
iron content. However, it should be noted that the BOS composi- SEM 1.5, 2.0 20, 30 28, 90 (days)
tion is highly variable depending on the raw materials, steel type, EDS 2.0 20 28 (days)
manufacturing process, and furnace conditions, among others. *
LVS: laboratory vane shear test; UC: unconfined compression test.

2.2. Sample preparation

The Tokuyama dredged clay was filtered through a 2 mm sieve

In order to obtain a comprehensive understanding the mechan-
ical and microstructural properties of the dredged marine clay sta-
bilized with BOS, a series of laboratory tests were carried out on
specimens with different water and BOS proportions, as presented
in Tables 4 and 5, and preparation procedure of sample was fol-
lowed by the Guidebook for Oceanographic Application of Con-
verter Slag [7]. The initial clay water content, w0, was
determined based on the clay liquid limit, wLL, and the sample con-
ditions in this study were set as 1.2, 1.5, 1.7, and 2.0 times the liq-
uid limit, considering the initial water contents of cement-treated
soil used in construction projects in coastal areas [34,39,41]. The
BOS content, BOSvol (%), which is the ratio of the BOS volume to
the total volumes, is defined as follows:
V BOS
BOSvol ¼  100 ð%Þ
V soil þ V water þ V BOS
where Vsoil and Vwater are the dried soil and water volumes, respec-
tively, and VBOS is the BOS volume.
100
Tokuyama Port clay
90
BOS < 5mm
80
Percent finer (%)
to remove coarse particles or other impurities. The BOS was air-
dried at a room temperature of 20 ± 3 °C for 1 d to dry only the
moisture of slag surface, i.e., the surface dry density condition.
The BOS water content following air-drying was 1.12%. Artificial
seawater with a 3.5% salinity was prepared as additional water
used in the desired water content of the experimental samples in
order to create identical conditions to those of marine clay in
coastal areas. The slurry obtained after adding BOS and additional
water to the marine dredged clay was fully mixed for 5 min using a
hand mixer.
LVS and UC tests were adopted to evaluate the strength devel-
opment at the early and later curing stages, respectively. A cylin-
drical mold of 60 mm in diameter and 60 mm in height was
prepared for the LVS test. A cylindrical tinplate mold with a diam-
ð1Þ
eter of 50 mm and a height of 100 mm was employed for the UC
test. The mixture was poured into the molds by dividing it into
three layers, and lightly tapped to remove the void of air bubbles
in the molds. The cylindrical molds were covered with polythene
wrap and cured under atmospheric pressure at a temperature of
20 ± 3 °C, while the cylindrical tinplate molds of the UC test were
cured in distilled water at room temperature. The specimen prepa-
ration for this procedure required approximately 30 min. Accord-
ingly, the curing starting time of curing was determined as
30 min following mixing.

70
For the sample preparation for the SEM and EDS, the stabilized
60 specimens listed in Table 3 were trimmed into small fragments
50 prior to drying in order to implement freeze-drying with ease.
40 The prepared fragments were frozen instantly by means of immer-
sion in liquid nitrogen of approximately 195 °C for 30 min. The
30
frozen samples were then evacuated at a pressure of 0.5 Pa at
20 40 °C for 1 d using vacuum drying apparatus. The freeze-dried
10 samples were broken up by finger pressure into pieces of approx-
0 imately 6 mm in diameter and 6 mm in height. The undisturbed
0.0001 0.001 0.01 0.1 1 10 surfaces of the broken samples were positioned upwards on an alu-
Particle diameter (mm) minum stub, and fixed using carbon bond and tape. Finally, the
samples on the aluminum stub were coated with gold in order to
Fig. 1. Particle-size distribution curves for Tokuyama Port clay and BOS. obtain a high-quality image from the SEM and EDS tests.
504 G. Kang et al. / Construction and Building Materials 203 (2019) 501–513

Table 5
Mixing proportions of BOS-treated soil in this study.

No. Volumetric Tokuyama Port clay BOS BOS- Density of BOS-

BOS treated treated soil (g/cm3)
Density Soil Water Added Total Total mass Normalized water Particle
addition soil
(g/cm3) particle water water and volume content, w0/wLL density = 3.15 g/cm3
rate, RBOS *
2.646 1.03 1.03* 1.03*
Case 1 20% Mass (g) 2372 2541 509 3050 5422 1.2 3038 8459 1.75
Volume (mL) 896 2467 494 2961 3857 964 4822
Case 2 20% Mass (g) 2042 2560 722 3282 5323 1.5 3117 8440 1.69
Volume (mL) 772 2485 701 3186 3958 989 4947
Case 3 20% Mass (g) 1907 2394 1080 3474 5382 1.7 3224 8606 1.66
Volume (mL) 721 2325 1049 3373 4094 1023 5117
Case 4 20% Mass (g) 1725 2239 1458 3697 5422 2.0 3340 8761 1.63
Volume (mL) 652 2173 1416 3589 4241 1060 5301
Case 5 30% Mass (g) 1939 2310 183 2493 4433 1.2 4257 8690 1.91
Volume (mL) 733 2243 178 2421 3154 1352 4505
Case 6 30% Mass (g) 1701 2114 620 2734 4435 1.5 4451 8886 1.88
Volume (mL) 643 2052 602 2654 3297 1413 4710
Case 7 30% Mass (g) 1565 2051 799 2850 4415 1.7 4534 8949 1.87
Volume (mL) 591 1991 776 2767 3359 1439 4798
Case 8 30% Mass (g) 1413 1869 1159 3028 4441 2.0 4690 9131 1.85
Volume (mL) 534 1815 1125 2940 3474 1489 4963
*
Specific gravity of artificial sea water.

2.3. Testing methods
In this study, all of the laboratory tests were conducted on
dredged marine clay stabilized with BOS to investigate the
geotechnical properties, strength development, and microstruc-
tural and mineralogical properties. A flow test was carried out on
all specimens immediately following mixing to determine the mix-
ture workability. The LVS test was conducted on samples during
the early curing stage (0.5–10 h), while the UC test was carried
out on samples during the later curing stages (10 h to 90 d). The
SEM and EDS analyses were conducted on eight samples and one
sample, respectively, in order to examine the reaction products
of the stabilized dredged marine clay, and SEM was additionally
carried out on two cement-treated clay samples and compared to
that of ordinary Portland cement.
2.3.1. Flow test
A flow test was carried out following the cylinder method of the
Japan Highway Public Corporation [11] in order to obtain the flow
value as a significant parameter in evaluating the fluidity and
workability for stabilized marine dredged clay. An open-end cylin-
drical acrylic mold with a diameter and height of 80 mm was ini-
tially placed on a metalized plastic plate and then filled with the
mixture up to the cylinder upper end so as to prevent overflowing
from the cylinder. The cylinder mold was lightly tapped on the
cylinder side with a finger so that the mixture surface was horizon-
tal and coincided with the cylinder upper end. The cylinder was
lifted vertically, allowing for the mixture to spread freely on the
plate, and the spread diameters in two perpendicular directions
were measured after 1 min. Finally, the flow value was determined
by the average value of spread diameters.
2.3.2. LVS and UC tests
The LVS test was performed to measure samples with compar-
atively low strengths and very early curing times, ranging from 0.5
to 10 h, when the samples could not stand on their own. The vane
diameter and height were both 20 mm, and a vane height/diameter
ratio (H/D = 1) was used. The rotation speed of the laboratory vane
was set at 6°/min, and the vane was rotated until 60°. Moreover,
the torque peaks in the mixtures were reached within 60° in all
experiments conducted.
The UC test was conducted following the Japan Industrial Stan-
dard JIS A 1216 [12] on the remaining samples at a later curing
stage, namely 10 h to 90 d, which could stand with sufficient
strength after completing the LVS test. A linear variable differential
transformer was used to measure the local internal strain, and the
compression strain rate was 1% per minute.
In this study, the undrained shear strength (su) obtained from
the LVS test was calculated using the equation qu = 2su. The
strength of the marine dredged clay stabilized with BOS was
expressed as either qu or 2su with the curing time.
2.3.3. SEM and EDS analysis
In order to investigate the changes in the matrix microstructure
of the stabilized clay qualitatively, SEM and EDS analysis for a min-
eralogical study were carried out using a JEOL JSM – IT300 SEM,
operating at 30 kV. The SEM was fitted with an energy-dispersive
X-ray spectrometer. By means of EDS, the specimen chemical com-
positions were analyzed. Micrographs were obtained using the
EDS2000 software.
3. Results and discussion
3.1. Physical properties of BOS-treated dredged clay
3.1.1. Stress-strain curve
Fig. 2 illustrates the stress-strain curves from 15 h to 90 d of
curing, obtained from the UC test on the samples with a BOS con-
tent of 30% and w0 = 1.5 wLL. As indicated in Fig. 2, the stress-strain
curve behavior changed significantly with the curing time. The
15 h curing curve exhibited a failure strain of approximately 2%
and was generally similar to the stress-strain property of general
soft clay. In the stress-strain curve after 7 d of curing, the axial
strain at failure was decreased, and was less than 2%. The axial
strain at failure in the sample with the longest curing period of
90 d was approximately 0.8%. These values are very similar to
those determined for cement-treated clay [22,20,8].
3.1.2. Flowability
Cement-treated and BOS-treated dredged clay can be utilized as
backfill, embankment, and sealing materials. These materials have
a specific layer thickness and gradient when placed in the con-
struction site. Moreover, the layer thickness and gradient are
determined by the construction conditions [34]. Therefore, the
flowability of a mixture immediately following mixing with a bin-
der is a very significant indicator for reflecting the workability at
 G. Kang et al. / Construction and Building Materials 203 (2019) 501–513 505

1600 350
Curing time BOS 20% less than 5mm
1400 BOS-treated Tokuyama Port Clay BOS 30% less than 5mm
90 days
w0= 1.5wLL 300 BOS 30% less than 2mm
1200 BOSvol = 30% Tokushima port clay (TMC)
TMC-bentonite mixture 1
Stress (kN/m2)

1000

Flowability, FA (mm)
28 days 250 TMC-bentonite mixture 2
Ministry of Transport (1999)
800

600
7 days
400
3 days
200 2 days 1 day
0 100
0 1 2 3 4 5
Axial strain (%)
Fig. 2. Stress-strain curves for BOS-treated Tokuyama Port clay (e.g. BOSvol = 30%
and w0 = 1.5wLL).
construction. Furthermore, the target flow value varies according
to the construction method and material application purpose.
Fig. 3 illustrates the variation in flowability (FA) with an increase
in the initial water content (w0/wLL) for all BOS mixtures in this
experiment. The flow value ranged from 80 to 98 mm, and the
flowability exhibited an increase with an increase in the initial
water content and a decrease with an increase in the BOS content.
It was found that the flowability increase rate between the initial
water contents of 1.2 and 2.0 w0/wLL demonstrated an approxi-
mately 18% increase, irrespective of the BOS content. Moreover,
the flowability decrease rate according to an increase in the BOS
content at the same initial water content ranged from 1.24 to
3.98%. These phenomena are predominantly owing to the changes
in the free-water content; that is, the free-water content increases
with an increase in the initial water content, and decreases with an
increase in fine aggregates of air-dried BOS, which absorb the free-
water content because of the increased specific surface.
Fig. 4 presents the relationship between the flowability and
undrained shear strength. Additional flow tests were carried out
for a mixture stabilized with BOS, where the BOSVol of less than
105
BOS 20%
BOS 30%
100
Flowability, FA (mm)
200
FA more than 100 mm
FA = 694.4·Su-0.32
150 R2 = 0.982
15 hours
6 7 8
50 FA less than 100 mm
FA = 148.98·Su-0.074
R2 = 0.832
0
10 100 1000 10000
Undrained shear strength, Su (Pa)
Fig. 4. Relationship between flowability and undrained shear strength.
2 mm diameter was 30%, untreated Tokuyama port clay (TPC),
and TPC-bentonite mixtures, and several data collected from the
literature [23] in order to assess the relationships between flowa-
bility and undrained shear strength accurately over a wide range.
These values are primarily governed by the specific surface area,
as well as the adsorption ability and void ratio related to the
inter-particle contact. For example, a large amount of cement-
treated dredged clay produced by the pre-mix vessel method has
frequently been used as construction material for back filling or
reclamation, and its flow value has conventionally been controlled
to be within the range of 100–150 mm, depending on the construc-
tion conditions [36]. Therefore, in this study, two ranges for the
relationship between the flowability and undrained shear strength
were proposed: flow values above and below 100 mm. The range of
flow values less than 100 mm was 80–98 mm, with the undrained
shear strength ranging from 406 to 4193 Pa. The flow values above
100 mm ranged from 103 to 280 mm, corresponding to undrained
shear strengths of 17–924 Pa. As illustrated in Fig. 4, the flowability
in the range of more than 100 mm decreased considerably with an
increase in the undrained shear strength. For less than 100 mm, the
flowability decreased gently with the undrained shear strength.
Two equations can be expressed with high determination coeffi-
cients (more than 100 mm: R2 = 0.98 and less than 100 mm:

R2 = 0.83), as follows:
95
FA ¼ 694:4  S0:32
u ðMore than 100 mmÞ ð2Þ
-3.92%
90 FA ¼ 149:0  S0:074 ðLess than 100 mmÞ: ð3Þ
u

We established that the undrained shear stress may be a signif-

85 -1.24% icant variable related to the flowability, which can be determined
using the proposed equations.
-2.97%
80 3.1.3. Strength development characteristics with curing time
-3.98% A series of LVS and UC tests were carried out to examine the
strength development characteristics of the dredged clay stabilized
75 with different BOS and initial water contents, from immediately
1.2 1.5 1.7 2.0 following mixing to long-term curing times varying from 0.5 h to
Initial water content, w0/wLL 90 d; the relationship between strength and curing time is plotted
in Fig. 5. The stabilized dredged clay strength development gener-
Fig. 3. Variation in flowability with initial water content for each BOS content. ally exhibited an increase with an increase in the BOS and decrease
506
10000
Inactive zone Active zone
1000
Slope 2: b2
2Su or qu (kPa)
100
Slope 1: b1
10
1 BOS 20%, 2.0 wLL
5 hours
0.1
0.1 1 10
Curing time (hours)
Fig. 5. Strength development against curing time on a log-log scale.
in the initial water content of the dredged clay. The strength devel-
opment was changed at a specific curing time with different
strength increment rates, which is the gradient for the relation of
the strength and curing time in the log-log scale graph. It was
observed that the strength development increased marginally up
to 5 h of curing time up to 5 h of curing time. The strength devel-
opment exhibited a reduction between 0.5 and 2 h of curing, even
though the curing time was passed, except for three samples,
namely BOSVol = 30%, and w0 = 1.2, 1.5, and 2.0 wLL. Moreover, the
strength increments of the excluded samples increased very
slightly with an increase in strength against the elapsed curing
time. This means that the strength did not develop within the
range of 0.5–2 h, and this period is known as a ‘‘dormant period”
in the concrete engineering field. Seng and Tanaka [29] reported
that the dormant period of marine clay treated with ordinary
Portland cement (OPC) is approximately 30 min immediately fol-
lowing mixing. It was found that the setting time at which the
strength development begins of dredged clay stabilized with BOS
is significantly longer than that of OPC; that is, the rate of the
hydration reaction increasing the stabilized dredged clay strength
Intensity (cps)
G. Kang et al. / Construction and Building Materials 203 (2019) 501–513
is very slow at the very early curing stage (within 5 h) compared to
Moderate zone OPC.
Fig. 6 presents the X-ray diffraction patterns of the BOS pro-
Slope 3: b3
vided in this study. The main minerals of BOS include calcium
hydroxide (Ca(OH)2), calcite (CaCO3), CaO, C2S, and wüstite (FeO).
The presence of FeO, srebrodolskite (Ca2Fe2O5), and fayalite (Fe2-
SiO4) indicates a higher iron oxide condition in BOS. In these X-
ray diffraction patterns, the C2S identified in the BOS was the major
mineral, in the form of b-C2S, which is present in cement, and this
mineral can be hydrated and gain strength owing to the hydration
72 hours reaction. Based on the results, it suggested that the stabilizing
mechanisms of the BOS-treated dredged clay are the hydration
BOS 20%, 1.2 wLL
BOS 20%, 1.5 wLL reaction of the C2S and the pozzolanic reaction of the Ca(OH)2
BOS 20%, 1.7 wLL and Si/Al obtained in the clay. Hence, possible reasons for the slow
BOS 30%, 1.2 wLL hydration rate are the deficiency of certain minerals (C3A or C4AF)
BOS 30%, 1.5 wLL that create a high hydration rate in very early curing, and the lack
BOS 30%, 1.7 wLL and low activity of C2S and C3S in the BOS compared to OPC
BOS 30%, 2.0 wLL
[25,26]. Wang et al. [40] mentioned another reason, namely that,
100 1000 10000 owing to the structure of the free lime in BOS being denser than
that of burnt lime and the decreased ability of moisture to react
with the free lime, the hydration reaction of free lime in the BOS
is slow compared to burnt lime, in which the hydration can be
completed within 30 min. We considered that this finding may
be useful in predicting or controlling the fluidity and the workabil-
ity of BOS-treated clay, depending on the construction conditions.
Moreover, after 5 h of curing, the strength development was shar-
ply increased with a high strength increment rate up to 3 d of cur-
ing governed in the hydration reaction. Thereafter, moderation
with curing periods (from 3 to 90 d) dominated by the pozzolanic
reaction was predominant, with a gentle slope, relative to a low
strength increment rate. Based on this behavior, we suggested that
the strength development can be divided into three zones in the
log-log scale graph: 1) inactive zone (0.5–5 h), 2) active zone (5 h
to 3 d), and 3) moderate zone (3–90 d). Fig. 7 illustrates the
strength development comparison between the BOS-treated
Tokuyama clay and cement-treated Tokuyama clay obtained in
the literature [13–15]. Data of the cement-treated Tokuyama clay
were obtained from samples of 10% and 20% cement content deter-
mined by the mass ratio, c* = mcement/(mclay + mcement). It has previ-
ously been reported that the strength development of cement-
treated clay increases linearly from immediately following mixing
and changes before and after 3 d of curing with different strength
increment rates. The strength values of the cement-treated clay
before 3 d of curing were higher than those of the BOS-treated clay,
1
1: Calcium hydroxide, Ca(OH)2
2: Calcite, CaCO3
3: Lime, CaO
4: Dicalcium silicate, Ca2SiO4
5: Wustite, FeO
6: Srebrodolskite, Ca2Fe2O5
7: Fayalite, Fe2SiO4
5
4 6
2
1
32 4
6
5 61
7 7 3
7 4 5
2 7
2 3
2-Theta (degrees)
Fig. 6. X-ray diffraction patterns of BOS used.
 G. Kang et al. / Construction and Building Materials 203 (2019) 501–513 507

100000 25
BOS 30%, 1.2 wLL
BOS 20%, 1.2 wLL
BOS 20%, 2.0 wLL
BOS 20%, 1.5 wLL
c* 20%, 1.5 wLL (Kang et al., 2016 and 2017a, b)
10000 c* 10%, 2.0 wLL (Kang et al., 2016 and 2017a, b) BOS 20%, 1.7 wLL
20 BOS 20%, 2.0 wLL
BOS 30%, 1.2 wLL
BOS 30%, 1.5 wLL

2Su at 0.5 hours /Y

1000 BOS 30%, 1.7 wLL
2Su or qu (kPa)

15 BOS 30%, 2.0 wLL

100 2Su at 0.5h/Y = 29.84·(w0/wLL)-3.533

R² = 0.944
10
10
72 hours

1 5
BOS
c* (OPC)
0.1 0
0.1 1 10 100 1000 10000
1 1.2 1.4 1.6 1.8 2 2.2
Curing time (hours)
Initial water content, w0/wLL
Fig. 7. Comparison of strength development between BOS-treated and cement-
treated Tokuyama Port clay. Fig. 8. Relationship between strength at curing of 0.5 h, normalized by volumetric
solid content, and initial water content.

while after 3 d of curing, the strength magnitudes of the BOS-

amount of free water and the void ratio. Based on the relationship
treated clay were larger or smaller than those of the cement-
illustrated in Fig. 7, the equation for estimating the strength at cur-
treated clay. Hence, we note that the strength increment rate with
ing of 0.5 h can be obtained, as per Eq. (6):
curing time may differ depending on the binder type.
- B
2Su0:5h ¼ Aðw0 =wLL Þ Y
3.2. Proposed strength estimation equations for each zone and its - 3:53
; ð6Þ
microstructure properties
2Su0:5h ¼ 29:84ðw0 =wLL Þ Y

where 2Su0.5h is the strength at curing of 0.5 h, w0/wLL is the initial

As mentioned previously, the strength development was classi-
fied into three zones, namely the inactive zone (0.5–5 h), the active
zone (5 h to 3 d), and the moderate zone (3–90 d), in accordance
with the strength increment rate. Consequently, equations for each
zone were proposed with various indices.
3.2.1. Strength estimation equation for inactive zone (0.5–5 h)
The strength at 0.5 h of curing and the strength increment rate
were used to propose the strength estimation equation for the
inactive zone. The strength at 0.5 h of curing was correlated with
water content, and A and B are the parameters related to the volu-
metric solid and initial water contents, respectively.
Fig. 9 illustrates the relationship between the strength incre-
ment coefficient in the inactive zone, b1, and the BOS mass ratio,
BOSmass. The b1 value generally increased with an increase in the
BOSmass except in the sample with BOSVol = 20% and w0/wLL = 1.7.
Moreover, it can be observed that the values of b1 in the BOS of
0.6
Strength increment coefficient, b1 (Δqu/Δt)

the volumetric solid content, Y, which is defined as follows: BOS 20%, 1.2 wLL
BOS 20%, 1.5 wLL
V BOS þ V Soil BOS 20%, 1.7 wLL
Y¼ ; ð4Þ 0.5
V BOS þ V Soil þ V Water BOS 20%, 2.0 wLL
BOS 30%, 1.2 wLL
where VBOS, VSoil, and VWater are the BOS, soil, and water volumes, BOS 30%, 1.5 wLL
respectively. 0.4 BOS 30%, 1.7 wLL
The strength increment rate (gradient for the relationship BOS 30%, 2.0 wLL
between the strength and curing time within 5 h), b1, obtained in
relation to the curing time strength and logarithm, was associated 0.3
with a BOS mass ratio. The BOS ratio is defined as the ratio of the b1 = 12.186·BOS2mass-
BOS mass to the mass of solid particles, as per Eq. (5): 13.654·BOSmass+3.853
mBOS 0.2 R² = 0.970
BOSmass ¼ ; ð5Þ
mBOS þ msoil
where mBOS and msoil are the BOS and soil dry masses, respectively. 0.1
Fig. 8 demonstrates the relationship between the strength at a
curing time of 0.5 h, normalized by the volumetric solid and initial
water contents. This relation exhibited a strong correlation with 0
the high determination coefficient, R2 = 0.944, and can be 0.5 0.6 0.7 0.8
expressed by a single line of a power function form. This relation Mass ratio of BOS slag, BOSmass
means that the factor governing the strength increase at curing
of 0.5 h is not a chemical effect such as hydration or pozzolanic Fig. 9. Relation of strength increment coefficient in inactive zone and BOS mass
reactions, but rather physical effects owing to a change in the ratio.
508 G. Kang et al. / Construction and Building Materials 203 (2019) 501–513

20% were smaller than those of the BOS of 30%. This relationship BOS 20%, 1.2 wLL
can be reasonably represented as a single line by the following 8 BOS 20%, 1.5 wLL
polynomial function: BOS 20%, 1.7 wLL
BOS 20%, 2.0 wLL

Strength at 1 h curing, ln(a1)

b1 ¼ ba  BOS2mass  bb  BOSmass þ bc
; ð7Þ
b1 ¼ 12:186  BOS2mass  13:654  BOSmass þ 3:853
where b1 is the strength increment coefficient in the inactive zone,
and ba, bb, and bc are the parameters depending on BOSmass.
An empirical equation for predicting the strength in the inactive
zone was proposed based on Eqs. (6) and (7), as follows:
B
2Su ðInactive zoneÞ ¼ Aðw0 =wLL Þ  Y ½1 þ b1 lnðt=0:5hÞ: ð8Þ
It should be noted that Eq. (8), composed using the volumetric
solid content and BOS mass ratio, could be estimated for the
strength in the inactive zone, ranging from 0.5 to 5 h.
3.2.2. Strength estimation equation for active zone (5 h to 3 d)
Kang et al. [13–15] suggested two equations that can estimate
the strength of cement-treated marine clay during early and later
curing times, based on the concept of changing the strength before
and after 3 d of curing in the log-log scale graph. The proposed
equations were composed of the strength at 1 h of curing and
strength increment coefficients (slope for the relationship between
the strength and curing time), and the 1 h of curing and strength
coefficients could be determined by the specific volume ratio and
cement content. In this study, the strength estimation equations
for the BOS-treated dredged clay in the active and moderate zones
were also proposed based on the strength at 1 h of curing and
strength increment coefficients.
Referring to Fig. 5, the strength development increased linearly
with the curing time on the log-log scale, and its relation can be
placed into a linear model by a power function for predicting the
strength of BOS-treated dredged clay in the active zone, for 5 h
to 3 d of curing, as expressed in Eq. (9):
lnðqu or 2Su Þ ¼ lna1 þ b2 lnðtÞ
expflnðqu or 2Su Þg ¼ expflna1 þ b2 lnðtÞg
; ð9Þ
qu or 2Su ¼ a1  expfb2  lnðtÞg
qu or 2Su ðactive zoneÞ ¼ a1  tb2
where qu or 2Su are the strengths obtained by the UC and LVS tests,
respectively, a1 is the strength at 1 h of curing, and b2 is the strength
increment coefficient, which represents the slope for the strength
and curing time relation in the active zone.
After determining the strength parameter, a1, using regression
analysis on the relationship between the strength and curing time
in the log–log scale, a1 was associated with a specific volume ratio
(v0 /vLL), which is the specific volume of the BOS-treated clay nor-
malized by that at the dredged clay liquid limit. The definitions
for the indices used are as follows:
V Soil þ V Water þ V BOS
v 0 ðBOS  treated clayÞ ¼ ; active zone, b2, and the BOS mass ratio with the cement-treated
V Soil þ V BOS
V þV
v LL ðDredged clayÞ ¼ Soil Water : ð10Þ
V Soil
Fig. 10 illustrates the relationship between the logarithms of a1
and those of the specific volume ratio, lnv0 /lnvLL, and compares this
with the result of the cement-treated Tokuyama clay obtained
from Kang et al. [14]. The a1 value exhibited a reduction with an
increase in the specific volume ratio for each BOS content, which
means an increase in the water content, and increased with an
increase in the BOS content. It was found that the a1 value of the
BOS-treated dredged clay was smaller than that of the cement-
treated dredged clay. The ln(a1) and lnv0 /lnvLL values exhibited
6 BOS 30%, 1.2 wLL
BOS 30%, 1.5 wLL
BOS 30%, 1.7 wLL
BOS 30%, 2.0 wLL
4 c*=10, 20, 30%, 1.5 wLL (Kang et al., 2017a)
c*=10, 20, 30%, 2.0 wLL (Kang et al., 2017a)
BOS BOS
2 30% 20%
OPC
c* = 10, 20, 30%
0
ln(a1) = 14.55-24.13(lnv'/lnvLL)
R² = 0.932 (BOS 30%)
-2
ln(a1) = 16.91-23.12(lnv'/lnvLL)
R² = 0.959 (BOS 20%)
-4
0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4
Normalized specific volume, lnv'/lnvLL
Fig. 10. Relationship between logarithms of a1 and logarithms of specific volume
ratio, lnv0 /lnvLL, and comparison with results of cement-treated Tokuyama clay.
strong correlations with the high determination coefficients, and
could be divided into two lines depending on the BOS contents.
The correlations were presented as expressed in Eq. (11):
lnða1 Þ ¼ c1  c2 ðlnv 0 =lnv LL Þ
or ; ð11Þ
0c2 =lnv LL
a1 ¼ expðc1 Þm
where c1 is the strength when the specific volume ratio is 0: lnv0 /
lnvLL = 0. Parameter c2 represents the strength increment rate with
the specific volume ratio (gradient of the relationship between the
logarithm of strength and the logarithm of the specific volume
ratio). The c1 and c2 values for each BOS content used were deter-
mined as illustrated in Eq. (11).
BOSvol ¼ 20% : a1 ¼ expð16:91Þm023:12=lnv LL
: ð110Þ
BOSvol ¼ 30% : a1 ¼ expð14:55Þm024:13=lnv LL
Based on the results for the relation of a1 and lnv0 /lnvLL, it was
observed that a1 can successfully be determined by the specific
volume ratio.
The strength increment coefficient, b2, in the active zone,
obtained from the graph for the relationship between the strength
and curing time in the log-log scale, exhibits a correlation with the
BOS mass ratio. The b2 data of the cement-treated clay calculated
by the cement mass ratio, mcement/(msoil + mcement), obtained from
Kang et al. [14] were added to the graph for comparison of the
strength increment rate according to the binder type. Fig. 11 plots
the relationship between the strength increment coefficient in the
clay data. It was found that the b2 of the BOS-treated dredged clay
was increased significantly with an increase in the binder mass
ratio. Moreover, the b2 values of the BOS of 20% exhibited a wide
range, in accordance with an increase in the binder mass ratio,
compared to that of the BOS of 30%. In the case of cement, b2 exhib-
ited a slight increase or constant value with an increase in the bin-
der mass ratio, irrespective of the cement and water contents. It
can be observed that the b2 of the BOS-treated clay was greater
than that of the cement-treated clay. Based on the results, it is
implied that the strength increment behavior owing to the hydra-
tion reaction that increases the strength in the active zone may dif-
fer in accordance with the binder type. Therefore, the correlation
 G. Kang et al. / Construction and Building Materials 203 (2019) 501–513 509

2.4 4
BOS 20%, 1.2 wLL BOS 20%, 1.2 wLL
BOS 20%, 1.5 wLL
Strength increment coefficient, b2 (Δqu/Δt)

BOS 20%, 1.5 wLL

Strength increment coefficient, b3 (Δqu/Δt)

2.2 BOS 20%, 1.7 wLL 3.5 BOS 20%, 1.7 wLL
BOS 20%, 2.0 wLL BOS 20%, 2.0 wLL
BOS 30%, 1.2 wLL BOS 30%, 1.2 wLL
2 3 BOS 30%, 1.5 wLL
BOS 30%, 1.5 wLL
BOS 30%, 1.7 wLL BOS 30%, 1.7 wLL
BOS 30%, 2.0 wLL 2.5 BOS 30%, 2.0 wLL
1.8 1.5 wLL, c*=10, 20, 30% (Kang et al., 2017a)
1.5 wLL, c*=10, 20, 30% (Kang et al., 2017a)
2.0 wLL, c*=10, 20, 30% (Kang et al., 2017a) 2.0 wLL, c*=10, 20, 30% (Kang et al., 2017a)
1.6 2 OPC
b2 = 3.31·BOFmass ratio - 0.96 BOS
c* =10, 20, 30% 20%
R2 = 0.755 (BOS 30%) 1.5
1.4
b2 = 4.53·BOFmass ratio - 1.35 BOF BOS
R2 = 0.977 (BOS 20%) 30% 30%
1.2 1
b3 = 8.33·BOSmass - 4.64
BOF
R2 = 0.976 (BOS 30%)
20% 0.5
1 OPC b3 = 0.99·BOSmass + 0.53
c*=10, 20, 30% R2 = 0.976 (BOS 20%)
0.8 0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
Mass ratio of BOS slag, BOSmass Mass ratio of BOS slag, BOSmass

Fig. 12. Relationship between strength increment coefficient in moderate zone and
Fig. 11. Relationship between strength increment coefficient in active zone and
BOS mass ratio with data of cement-treated Tokuyama Port clay.
BOS mass ratio with data of cement-treated Tokuyama clay.

the long-term curing time was insignificant. However, the b3 of

between b2 and the BOS mass ratio can be classified into BOS con-
the cement-treated clay increased linearly with the cement mass
tents and expressed as per Eq. (12):
ratio, except for c* = 30% and 1.5wLL. It was found that the b3 of
b2 ¼ d1  BOSmass  d2 the cement content of 10% was smaller than that of the BOS of
BOSvol ¼ 20% : b2 ¼ 4:53  BOSmass  1:35 ; ð12Þ 20% and 30%, while the cement-treated clay range was wider than
BOSvol ¼ 30% : b2 ¼ 3:31  BOSmass  0:96 that of the BOS. Based on the results, the correlation of b3 and the
BOS mass ratio can be expressed by a linear function, as expressed
where d1 is the strength increment rate with the BOS mass ratio and in Eq. (15):
d2 is the strength increment coefficient when mBOS/(msoil + mBOS) is
0. b3 ¼ e1  BOSmass þ e2
By substituting Eq. (11), determining the strength at 1 h of cur- BOSvol ¼ 20% : b3 ¼ 0:99  BOSmass þ 0:53 ; ð15Þ
ing, and Eq. (12), predicting the strength increment coefficient in BOSvol ¼ 30% : b3 ¼ 8:33  BOSmass  4:64
the active zone, into Eq. (9) regarding the strength estimation
model of the active zone, the empirical equation for estimating where e1 is the strength increment rate with the BOS mass ratio and
the strength of the BOS-treated dredged clay in the active zone e2 is the strength increment coefficient when mBOS/(msoil + mBOS) is
can be developed as follows: 0.
By substituting the strength estimation equation for the active
qu or 2Su ðactive zoneÞ ¼ expðc1 Þm0c2 =lnv LL  t ðd1 BOSmass d2 Þ
: ð13Þ
zone, Eq. (13), and the strength increment coefficient for the mod-
The proposed Eq. (13) can estimate the strength in the active
zone, at 5 h to 3 d, using the specific volume ratio and BOS mass 2000
BOS 20%, 1.2 wLL
ratio. BOS 20%, 1.5 wLL
BOS 20%, 1.7 wLL
Calculated strengths, 2Su or qu (kPa)

3.2.3. Strength estimation equation for moderate zone (3–90 d) BOS 20%, 2.0 wLL
BOS 30%, 1.2 wLL
For the moderate zone (3–90 d), we proposed the strength esti- 1500 BOS 30%, 1.5 wLL
mation model as per Eq. (14): BOS 30%, 1.7 wLL
 BOS 30%, 2.0 wLL
qu ðmoderate zoneÞ ¼ quð3 daysÞ 1 þ b3 lnðt=t 3 days Þ ; ð14Þ 1:1

where qu (moderate zone) is the strength, b3 is the strength incre- 1000

ment coefficient in the moderate zone, which is the gradient for
the relationship between the strength and curing time, and t and t3-
days are the curing time and 3 d of curing time, respectively.
The b3 value was determined by means of regression analysis
for the relationship between the strength and curing time loga- 500
rithm in the semi-log scale graph, and correlated by the BOS mass
ratio. Fig. 12 illustrates the relation of the b3 and BOS mass ratio,
and a comparison with that of Portland cement obtained from
the literature [14]. It was revealed that the b3 value of the BOS of 0
30% increased marginally with an increase in the BOS mass ratio, 0 500 1000 1500 2000
while the b3 of the BOS of 20% was very slight with a variation in Measured strengths, 2Su or qu (kPa)
the BOS mass ratio. This behavior means that, for the BOS of 20%,
the strength development rate owing to pozzolanic reaction at Fig. 13. Comparison of calculated and measured strengths.
510 G. Kang et al. / Construction and Building Materials 203 (2019) 501–513

Type II C-S-H

Type II C-S-H
Ettringite
Clay
matrix

AFm phase
AFm
phase

(a) w' = 1.5wLL, BOSVol. = 20%, T = 28days (b) w' = 1.5wLL, BOSVol. = 20%, T = 90days

Type II C-S-H

BOS slag

AFm
phase
Type II C-S-H
Clay
matrix

AFm phase

(c) w' = 1.5wLL, BOSVol. = 30%, T = 28days (d) w' = 1.5wLL, BOSVol. = 30%, T = 90days

Ettringite
cluster

Activated
BOS slag

Type II C-S-H

(e) w' = 2.0wLL, BOSVol. = 20%, T = 28days (f) w' = 2.0wLL, BOSVol. = 20%, T = 90days

Clay
matrix

Type II C-S-H

Type II C-S-H
AFm phase

AFm phase

Clay Clay
matrix matrix

(g) w' = 2.0wLL, BOSVol. = 30%, T = 28days (h) w' = 2.0wLL, BOSVol. = 30%, T = 90days
Fig. 14. SEM micrograph of BOS-treated Tokuyama port clay.
 G. Kang et al. / Construction and Building Materials 203 (2019) 501–513 511

erate zone, Eq. (15), into the proposed Eq. (14), the equation for
predicting the strength in the moderate zone was proposed, as
follows:
qu ðmoderate zoneÞ ¼ expðc1 Þm0c2 =lnv LL  t 31days
d ðBOSmass Þd2
 
 1 þ b3 lnðt=t 3 days Þ :
Fig. 13 provides a comparison of the measured strengths and
those calculated by the empirical equations proposed in this study.
The measured and calculated values agreed strongly with the over-
all data.
It should be noted that this study was focused on understanding
the characteristics of the strength behavior of BOS-treated dredged
clay at the very early and later curing times, and a comparison with
that of Portland cement-treated dredged clay. However, the
strength behavior of the BOS-treated dredged clay may be affected
by the BOS size, free-CaO content, and soil types, among other fac-
tors. Therefore, it is essential to investigate the parameter varia-
ð16Þ tions of the proposed equation with various considerations, such
as the BOS properties and soil types.
3.2.4. SEM and EDS analysis
In general, the cement stabilization process involves the growth
of newly formed phases within soil textiles by means of the hydra-
tion reaction of clinker and pozzolanic reactions in calcium

Ettringite
cluster

Type II C-S-H
Type II C-S-H

Type II C-S-H

AFm phase

(a) w' = 1.5wLL, cement content = 10%, T = (b) w' = 1.5wLL, cement content = 20%, T =
28days 28days

Fig. 15. SEM micrographs of cement-treated Tokuyama Port clay.

10000 times

Fig. 16. EDS analysis of small particles on activated BOS.

512 G. Kang et al. / Construction and Building Materials 203 (2019) 501–513
hydroxide (Ca(OH)2) and silica/alumina from soil. New cementi-
tious compounds in the soil fabric result in hardening occurring,
owing to binding to the soil microstructure, thereby increasing
the inter-particle contacts and filling the pore spaces in the
mixture. Furthermore, the geotechnical characteristics such as per-
meability, strength, and stiffness can be modified by the stabiliza-
tion reaction. Hence, the factors affecting the formation of
cementitious compounds, such as the clinker amount, degree of
alkaline dissolution of the alumina and silica in soil, pozzolanic
sources, diffusion, and surficial aqueous complexation, are signifi-
cant in the stabilization process [28,42]. We considered that the
stabilization mechanisms of the BOS-treated clay are the hydration
and pozzolanic reactions, because BOS also contains mineral com-
pounds similar to those of Portland cement and free lime (F-CaO),
which cause the hydration and pozzolanic reactions to occur.
Fig. 14 presents the microstructural aspects of the Tokuyama clay
stabilized with BOS contents of 20 and 30%, and initial water con-
tents of 1.5 and 2.0wLL, following 28 and 90 d of curing. The retic-
ulation structure of the amorphous C-S-H gel and platy AFm phases
on all samples were observed with intergrowths of rod-like ettrin-
gite within the flat clay structure and flocculated clay-BOS cluster.
The degree of cementitious compounds, reticulation of C-S-H gel,
AFm phases, ettringite, and flocculated clay-BOS clusters were
more apparent with an increase in the BOS content and curing
time. In Fig. 14(c) and (d), it can be observed that the BOS particles
were wrapped by C-S-H gels, AFm, and flocculated clay-BOS clus-
ters. In Fig. 14(c), it can be observed that the BOS particles retained
their smooth surfaces, which means that the hydration reaction
was rarely developed. However, in Fig. 14(e), it can be observed
that a dense matting of amorphous C-S-H gel and small particles
covered the BOS surface. This means that the hydration reaction
occurred from the BOS. From the micrographs in Fig. 14(e), it
was determined that the BOS is a highly reactive material that
has the potential to undergo hydraulic reactions with clay clusters.
Fig. 15 illustrates the SEM micrograph of the cement-treated
Tokuyama Port clay. The C-S-H and AFm forms are commonly
observed with long needle-type ettringite in the cement-treated
sample. The sample of cement content of 10% exhibited C-S-H gel
and ettringite with large openings, and the void among the cemen-
titious compounds was significantly larger than those of the BOS-
treated clay. In case of the cement content of 20%, it can be seen
that the reticulation structure of the C-S-H gel and platy AFm
phases became more evident with an increase in the cement con-
tent, and denser than those of the BOS-treated clay. Fig. 16 pre-
sents the typical EDS analysis with the identified chemical
elements of the small particles in the activated BOS of Fig. 14(e).
From the EDS analysis results in the yellow area of the SEM micro-
graphs, magnified 10,000 times, the dominant chemical elements
are oxygen (O) and calcium (Ca). Therefore, it is considered that
the small particles on the BOS surface were calcium oxide (CaO),
which was found to exist around the BOS.
4. Conclusions
An investigation into the strength development and microstruc-
tural properties of BOS-treated dredged clay has been presented in
this paper. Based on the results of the strength and SEM micro-
graph analyses in this study, the following conclusions are
summarized:
1. The flow value exhibited a strong correlation with the
undrained shear strength, which may be a significant variable
for determining the flowability of the BOS-treated dredged clay.
2. A dormant period was observed, in which the strength of the
BOS-treated clay did not increase between the 0.5 and 2 h
curing times. The strength development can be classified into
three zones, according to the strength increment rates; 1) inac-
tive zone (0.5–5 h), 2) active zone (5 h to 3 d), and 3) moderate
zone (3–90 d).
3. For the inactive zone, the equation for estimating the strength
was proposed according to the volumetric solid content, initial
water content, and strength increment rate. The strength in the
active and moderate zones could be estimated based on the
equation consisting of the normalized specific volume (v0 /vLL)
and strength increment rates for each of the active and moder-
ate zones.
4. The strength increment rates of the BOS changed according to
the active and moderate zones. Moreover, the value and range
of the strength increment rate varied with the binder type,
BOS, and cement. For the active zone, the BOS strength incre-
ment rate increased significantly with an increase in the BOS
mass due to the hydration reaction. The strength increment rate
behavior in the moderate zones differed with the change in the
BOS content. Based on this result, the effect of the pozzolanic
reaction in increasing the strength at the long-term curing time
can be changed with the BOS content.
5. The reticulation structures of the amorphous C-S-H gel and
platy AFm phases on all samples were observed with rod-like
intergrowths. The degree of cementitious compounds became
more evident with an increase in the BOS content and curing
time. The C-S-H gel and ettringite with large openings, and
the voids among the cementitious compounds of the BOS-
treated clay, differed compared to cement-treated clay.
Conflict of Interest
None.
Acknowledgements
The authors gratefully acknowledge the financial support
provided by the JFE Steel Corporation. This work was supported
by the National Research Foundation of Korea (NRF) grant, funded
by the Korean government (MSIT) (NRF-2018R1C1B6008095). We
also wish to thank Dr. Sugiyama, Yuri and Dr. Kaneko, Takashi of
the Port and Airport Research Institute in Japan for their support
in the SEM.
References
[1] Act on the Prevention of Marine Pollution and Marine Accidents in Article 5,
Section 1. 1973. (in Japanese).
[2] P. Ahmedzade, B. Sengoz, Evaluation of steel slag coarse aggregate in hot mix
asphalt concrete, J. Hazard. Mater. 165 (2009) 300–305.
[3] I. Arribas, I. Vegas, J.T. San-Jose, J.M. Manso, Durability studies on steelmaking
slag concretes, Mater. Des. 63 (2014) 168–176.
[4] I.M. Asi, Evaluating skid resistance of different asphalt concrete mixes, Build.
Environ. 42 (2007) 325–329.
[5] E. Belhadj, C. Diliberto, A. Lecomte, Characterization and activation of Basic
Oxygen Furnace slag, Cem. Concr. Compos. 34 (2012) 34–40.
[6] Y.F. Deng, T.W. Zhang, Y. Zhao, Q.W. Liu, Q. Wang, Mechanical behavior and
microstructure of steel slag-based composite and its application for soft clay
stabilization, Eur. J. Environ. Civ. Eng. 21 (2017) 1–16.
[7] Guidebook for Oceanographic Application of Converter Slag, 2008 (in
Japanese).
[8] S. Horpibulsk, R. Rachan, A. Suddeepong, A. Chinkulkijniwat, Strength
development in cement admixed Bongkok clay: laboratory and field
investigations. 51 (2), 239–251.
[9] K. Horii, N. Tsutsumi, Y. Kitano, T. Kato, Processing and reusing technologies for
steelmaking slag. Nippon Steel Technical Report. 104, 123–129.
[10] JGS 0211 2009. Test method for pH of suspended soils (in Japanese).
[11] JHS A 313 1992. Cylinder method for consistency test, Japan Road Association
(in Japanese).
[12] JIS A 1216 2008. Method for Unconfined Compression Test of Soils (in
Japanese).
 G. Kang et al. / Construction and Building Materials 203 (2019) 501–513
[13] G.-O. Kang, T. Tsuchida, A.M.R.G. Athapaththu, Engineering behavior of
cement-treated marine dredged clay during early and later stages of curing,
Eng. Geol. 209 (2016) 163–174.
[14] G.-O. Kang, T. Tsuchida, Y.-S. Kim, Strength and stiffness of cement-treated
marine dredged clay at various curing stages, Constr. Build. Mater. 132 (2017)
71–84.
[15] G.-O. Kang, T. Tsuchida, Y.-S. Kim, W.-J. Baek, Influence of humic acid on the
strength behavior of cement-treated clay during various curing stages, J.
Mater. Civ. Eng. 29 (8) (2017) 8.
[16] H. Kinoshita, K. Ichii, H. Morikawa, H. Takahashi, H. Shinozaki, Y. Takahashi,
Characteristics and assessment of seismic behavior of improved ground by
sand compaction pile method using iron and steel slag as base of gravity
caisson structure, Jpn. Geotech. J. 7 (1) (2012) 323–337 (in Japanese).
[17] M. Kitazume, M. Terashi, The Deep Mixing Method, CRC Press Taylor & Francis
Group, London, UK, 2013.
[18] A.R. Lee, Blastfurnace and Steel Slag: Production, Properties and Uses, Edward
Arnold, London, 1974.
[19] S. Liu, Z. Wang, X. Li, Long-term properties of concrete containing ground
granulated blast furnace slag and steel slag, Mag. Concr. Res. 66 (21) (2014)
1095–1103.
[20] G.A. Lorenzo, D. Bergado, Fundamental parameters of cement-admixed clay–
new approach, J. Geotech. Geoenviron. Eng. 130 (10) (2004) 1041–1050.
[21] P.Y. Mahieux, J.E. Aubert, G. Escadeillas, Utilization of weathered basic oxygen
furnace slag in the production of hydraulic road binders, Constr. Build. Mater.
23 (2009) 742–747.
[22] N. Miura, S. Horpibulsk, T.S. Nagaraj, Engineering behavior of cement
stabilized clay at high water content, Soils Found. 41 (5) (2001) 33–45.
[23] Ministry of Transport, The Fifth District Port Construction Bureau, 1999.
Pneumatic flow mixing method (in Japanese).
[24] M. Pasetto, N. Baldo, Experimental evaluation of high performance base course
and road base asphalt concrete with electric arc furnace steel slags, J. Hazard.
Mater. 181 (2010) 938–948.
[25] H.Y. Poh, G.S. Ghataora, N. Chazireh, Soil stabilization using basic oxygen steel
slag fines, J. Mater. Civ. Eng. 18 (2) (2006) 229–240.
[26] W. Qiang, S. Mengxiao, Y. Jun, Influence of classified steel slag with particle
sizes smaller than 20 mm on the properties of cement and concrete, Constr.
Build. Mater. 123 (2016) 601–610.
[27] L. Rondi, G. Bregoli, S. Sorlini, L. Cominoli, C. Collivignarelli, G. Plizzari,
Concrete with EAF steel slag as aggregate: a comprehensive technical and
environmental characterization, Compos. B 90 (2016) 195–202.
513
[28] K.L. Scrivener, The microstructure of concrete, in: J.P. Skalny (Ed.), Materials
Science and Concrete, American Ceramic Society, Westerville, OH, USA, 1991,
pp. 127–161.
[29] S. Seng, H. Tanaka, Properties of cement-treated soils during initial curing
stages, Soils Found. 51 (5) (2011) 775–784.
[30] D.H. Shen, C.M. Wu, J.C. Du, Laboratory investigation of basic oxygen furnace
slag for substitution of aggregate in porous asphalt mixture, Constr. Build.
Mater. 23 (2009) 453–461.
[31] W.G. Shen, M.K. Zhou, W. Ma, J.Q. Hu, Z. Cai, Investigation on the application of
steel slag–fly ash–phosphogypsum solidified material as road base material, J.
Hazard. Mater. 164 (1) (2009) 99–104.
[32] C.J. Shi, J.S. Qian, High performance cementing materials from industrial slags–
a review, Resour. Conserv. Recycling 29 (2000) 195–207.
[33] C. Shi, Steel slag–its production, processing, characteristics, and cementitious
properties, J. Mater. Civ. Eng., ASCE 16 (2004) 230–236.
[34] Y.X. Tang, Y. Miyazaki, T. Tsuchida, Practices of reused dredging by cement
treatment, Soils Found. 41 (5) (2001) 129–143.
[35] H. Takahashi, Y. Morikawa, H. Shinozaki, H. Kinoshita, K. Maruyama, Ground
stability against backfill loading of SCP improved ground using solidified iron–
and–steel slag, Jpn. Geotech. J. 6 (1) (2011) 81–95 (in Japanese).
[36] Technical Manual of Pre-Mix Vessels Method, 2013. Association of Pre-Mix
Vessels Method (in Japanese).
[37] The Japan Iron and Steel Federation: Guidebook for Oceanographic Application
of Converter Slag. 2008 (in Japanese).
[38] Committee for Utilization of Slag and Expert Committee on Steel Slag:
Tentative Method of Analysis of Free Lime in converter Slag and The Japan Iron
and Steel Federation 1980 Electric Furnace Slag.
[39] T. Tsuchida, Y.X. Tang, Y. Watabe, Mechanical properties of lightweight treated
soil cured in water pressure, Soils Found. 47 (4) (2007) 731–748.
[40] G. Wang, Y. Wang, Z. Gao, Use of steel slag as a granular material: Volume
expansion prediction and usability criteria, J. Hazard. Mater. 184 (2010) 555–
560.
[41] Y. Watabe, T. Noguchi, Site-investigation and geotechnical design of D-runway
construction in Tokyo Haneda Airport, Soils Found. 51 (6) (2011) 1003–1018.
[42] A. Wilkinson, A. Haque, J. Kodikara, Stabilisation of clayey soils with industrial
by-products: part A, Ground Improvement 163 (G13) (2010) 149–163.
[43] Y. Xue, S. Wu, H. Hou, J. Zha, Experimental investigation of basic oxygen
furnace slag used as aggregate in asphalt mixture, J. Hazard. Mater. 138 (16)
(2006) 261–268.